Although the engineer does not have to understand Philosophy to do engineering, if he wishes to engage in the technological forecasting and assessment process he must be able to relate what he does to the rest of society -- in this case to grasp the underlying philosophy.  The principles and methodology cannot be divorced from the plans, policies, and decisions of others.  The engineer must be able to communicate his principles and methods -- to economists, to sociologists, to planners, and so on.

It is in precisely this area that engineers are having trouble, for such understanding and communication require comprehension of the philosophies that underlie the different specialties and disciplines.  If an engineer is to communicate successfully with, say, an economist, he must understand what it means that the economist will tend to use Lockean methods in forecasting whereas the engineer has been trained (primarily) as a Leibnizian.

Moreover, the effective application of the various methodologies of technological forecasting and assessment requires understanding the philosophical underpinnings of these methodologies.  It is because such understanding is so widely lacking that trend analysis, Delphi techniques, and the other forecasting methods to be discussed are often misapplied.

No matter how well established the field of technological forecasting becomes, it can never be a purely technical or scientific concern.  It will always retain a basic philosophical element expressed by any one of the following questions: What permits us to extrapolate from the past or present to the future?  What
guarantees are there that the future will behave like ether present or past?  What firm assurance do we have that the future will behave as our projections (i.e. our models) of it forecast (i.e. predict)?

Whatever approach we take toward answering these questions, our answers will be indicative of a basic philosophical stance, indeed, of a basic image of reality.   Yet not only are we generally unaware of the different philosophical images that underlie our various technical models, but each of us has a fundamental
image of reality that runs so deep that often we are the last to know that we hold it.  As a result, we disagree with our fellows and we experience inner conflicts without really knowing why.  What's worse -- we ensure this ignorance by hiding behind catchwords and fancy names for techniques.  The field of
technological forecasting and assessment is no less remiss than many other disciplines in this respect.

One of the basic purposes of this article is to underscore these fundamental differences and conflicts of methodology so that hopefully one can be in a better position to choose explicitly a desired approach.

To accomplish this objective we shall consider some of the more significant philosophical stances taken toward the problem of predicting the future, or, more precisely, toward defining the criteria that would "guarantee" our extrapolation from the present to the future.  Also, we shall show that each of these stances differs sharply from the others and that each has its strengths as well as weaknesses.  This awareness constitutes a strength.  To show that there is no one mode of extrapolation that can satisfy our every requirement -- that there is no one mode that is best in all senses and for all circumstances -- is not to say that each of these modes is not uniquely or better suited for some special set of circumstances.

Bear in mind as we proceed that the question of concern is not how we can know the future with perfect certainty; here the answer clearly is that we can't.  However, neither can we know all there is to know about the present with perfect certainty.  The real question is what we can know of the future, and, even more to the point, how we can prove it.  It is on this very issue that the difference between these philosophies, or systems of inquiry, arises.

Leibnizian inquiry systems
The first philosophy to be discussed is that which underlies the major part of theoretical science -- the philosophy of Leibniz.  The sense of Leibnizian inquiry can be captured rather quickly and generally in terms of the following characteristics: Truth is analytic; i.e., the truth content of a system is associated entirely with its formal content.  A model of a system is a formal model and the truth of the model is measured in terms of its ability to offer a theoretical explanation of a wide range of general phenomena and in our ability as model-builders to state clearly the formal conditions under which the model holds.  A corollary to this is that the truth of the model does not rest upon any external considerations; in other words, the model is independent of the raw data of the external world.

In short, Leibnizian inquiry systems are the epitome of formal, symbolic systems.  They will characteristically strive to reduce any problem to a formal mathematical or symbolic representation.  The guarantor of such systems traditionally has been the precise specification of what shall count as a proof for a derived theorem or proposition; other guarantor notions include those of internal consistency, completeness, and comprehensives.

The laws of physics are examples of Leibnizian truths.  Also, computer simulation models of industrial processes are typical of the types of Leibnizian approaches that have been comparatively successful as forecasting tools in industry.  One can often model a proposed plant to a sufficient degree to examine alternative configurations before investments are made.

A prime example of Leibnizian inquiry is the field of operations research (OR), in the sense that the major energies of the profession have been almost exclusively directed toward constructing and exploring highly sophisticated formal models.  OR is the prime example of Leibnizian inquiry not because there is no utilization of external data whatsoever in OR models, but because much more attention is paid to teaching students of OR how to build sophisticated models that in teaching them equally sophisticated methods of data collection and analysis.

Two specific Leibnizian approaches to technological forecasting are correlation analysis is and substitution analysis.  Both result from an analogy with classical growth models governing such biological phenomena as the growth of cells and the growth of species.  The analogy assumes that, like biological phenomena, technological development passes through some fundamental, characteristic phases, such as birth, growth, and death.  For example, the process that governs the rate of transfer of technology can be represented as a diffusion-of-information process that is very close to the kind of diffusion a biological organism goes through in searching for food.  As a result of this kind of model, one may infer that certain curves related to the growth of technology are correlated.  A common example is the use of the performance of military aircraft at some point in time to infer the performance of civilian aircraft at some later time.  The model is predicting the time it will take military technology to diffuse into the civilian market.

The substitution curve analysis also employs the growth analogy, but in quite a different way.  The same kind of curve that is characteristically used to describe the growth of a biological population in a space of finite resources is also used to describe the percentage of the market that a new technology has and will assume.  Substitution analyses are characteristically given in the form of curves or tables indicating the percentage of substitution that has taken place by a new technology in a certain market in various fixed time periods.  The rule of thumb on the part of those who utilize the technique for planning is that by the time the process has reached a substitution level of 15 percent, it is usually irreversible ant the resulting forecasting curve is a useful projector of things to come.

Typical substitution curves have been exhibited for such transfers as man-made fiber to natural fibers; water-based paints to oil paints; man-made flooring to all flooring; synthetic rubber to natural rubber; margarine to butter; and vacuum tubes to transistors or transistors to integrated circuits.
 

The Leibnizian character of these models can be illustrated rather easily by spelling out a number of assumptions that underlie their applicability.  These assumptions are usually implicit.  For one, it seems to be an implicit assumption that such forecasts can be relied on to predict the future because the models reveal or embody a fundamental, enduring, structural feature of reality; e.g., the supposed basic features that govern the growth of biological phenomena.  A second assumption is that the models can be widely applied, again because they supposedly embody a characteristic process that underlies a wide range of technical and social processes.  In other words, the assumption is not only that a wide range of processes can be described in terms of these models but that the models actually underlie the behavior of a large number of processes; i.e., that in some sense the models are real.  In this sense, the most fundamental unspoken assumption is that as characteristic features of reality the models make possible the date that are fitted to them; the data do not make possible the models.  Indeed, the models implicitly assume that for a wide range of phenomena, there can be found the "right kind of data" that will fit the models; hence, their universal applicability is perpetually assured.  In this sense, the models take on the tenor of self-fulfilling prophecies.

For which problem situations are Leibnizian analyses most appropriate?  First, the situations must be so simple and well understood that they can be modeled.  Thus Leibnizian inquiry is best suited to definable, well-structured problems for which there exist an analytic formulation and solution.  Second, the modeler must have strong reasons for believing in the assumptions that underlie Leibnizian inquiry.  In a basic sense, the fundamental guarantor of Leibnizian inquiry is the understanding of the model-builder; he must understand the situation completely to believe he has represented it "accurately" and "faithfully".

The abuse of the technique usually occurs when there is not a good understanding (or no attempt to arrive at such) of the particular causal model underlying a particular correlation or substitution result.  Without comprehending the relationships in the model that produce the predicted effect, there is always the danger that a sudden change in the nature of the model will invalidate the ability to utilize the projections.  For example, the beginning substitution of plastics for metals in cars, as estimated by some forecasters, may be affected strongly, or even reversed in direction, by the recent shift in emphasis on safety.

There is no way in the correlation or substitution analyses to predict specific technological breakthroughs.  Therefore, all predictions hold only until a new technology or new synthesis of technology appears on the scene to begin a new set of curves.  For example, predictions based upon core memory technology for computers will not necessarily predict the effect of introducing bubble memory technology.  (However, once bubble memories are on the market a substitution process may become observable.)
 

Lockean inquiry systems
Lockean philosophy underlies the major part of empirical science, and its sense can be rather quickly and generally grasped in terms of the following characteristics: Truth is experiential; the truth content of a system is associated entirely with its empirical content.  A model of a system is an empirical model an its truth is measured in terms of our ability (1) to reduce every complex proposition to its simple empirical referents (simple observations) and (2) to ensure the validity of each of the simple referents by means of the widespread, freely obtained agreement between different human observers..

A corollary is that the truth of the model does not rest upon the prior assumption of any theory.  The only general propositions that are accepted are those justified through direct observation.

Lockean inquiry systems are the epitome of experimental, consensual systems.  They start form a set of elementary empirical judgments ("raw data," observations, sensations) and build up a network of ever expanding, increasingly more general factual propositions.  Whereas in Leibnizian inquiry the networks are theoretically, deductively derived, in a Lockean system they are empirically, inductively derived.  The guarantor of such systems has traditionally been the function of human agreement -- an empirical generalization is judged "objective," "true," or "factual" if there is sufficient widespread agreement on it by a group of "experts."  The final information content of a Lockean system is identified almost exclusively with its empirical content.

Statistics provide a good example of Lockean methodology.  In statistics the data vote, in a sense, on their own degree of validity in terms of probabilities, correlation coefficients, confidence limits, variances, etc.  A human may then judge if the degree of validity is sufficient to infer a prediction. Pure experimentation, in the sense of measuring phenomena, is a typical Lockean endeavor.  Many of the current generation of predictive economic models are basically Lockean in nature, since they rest largely on regression analyses of historical data.

In technological forecasting, trend extrapolation and regression analysis are simple and common examples of Lockean inquiry.  In the typical application of trend extrapolation, the performance over time of various technological indicators (e.g., computer speed, aircraft carrying capacity, material strength, energy production) is plotted and then the curves are extrapolated to give future trends.

Even where the curve extrapolation procedure is governed by complex mathematical considerations, the process still essentially Lockean.  The reason is that except for the possibility of statistical considerations, no theoretical model of the underlying phenomenon is used to guide the collection of the initial data or subsequent analysis, and, in this case, the extrapolation procedure.  In other words, the activities of theoretical explanation or justification, raw data collection, and curve extrapolation are assumed to be separable or independent of one another.  However, in a fundamental sense this is not, and never can be, the case.  They may not be related by an explicit well-developed formal theory, but they are related nonetheless.  One cannot consistently maintain that one can know very little of what the future will be like, and then argue that one knows with confidence that such and such a data set is a "relevant" and "reasonable" data base upon which to base a projection of what the future will be like.  The point is that to make the judgment that a particular data set is relevant to a projection of the future is to articulate a theory -- at the very  least, a point of view -- with respect to what the future will be like.

A more recent and far more interesting example of Lockean inquiry is the Delphi technique, first pioneered by Dalkey, Helmer, and Rescher at RAND.  In very simple terms, Delphi is a procedure for fostering a communication process among a large group of individuals.  In assessing the potential development of a technical area, a large group (typically in the tens or hundreds) is asked to "vote" on when they think certain events will occur.  A major premise underlying the approach is the assumption that a large number of expert judgments is required to treat any issue adequately.  (A face-to-face exchange among the group members would be inefficient or impossible because of the cost and time in bringing them together.)

The Delphi procedure is about as pure and perfect a Lockean procedure as one could hop to find.  For one, the "raw data inputs" are the opinions or judgments of the experts.   For another, the validity of the resulting judgment of the entire group is typically measured in terms of the explicit degree of consensus among the experts.  The feature that serves to distinguish the Delphi from and ordinary polling procedure is the feedback of the information gathered from the group and the opportunity of the individuals to modify or refine their judgments based upon their reaction to the collective views of the group.  Secondary characteristics are various degrees of anonymity enjoyed by the individual, and collective responses that avoid undesirable psychological effects with respect to the individual participant.

The strength of Lockean inquiry lies in its ability to sweep in rich sources of experimental data.  Indeed, the sources are so rich that they literally overwhelm the current analytical capabilities of most Leibnizian systems.  The weaknesses are those that beset all empirical systems.  Although experience is undoubtedly rich, it can also be extremely fallible and misleading.  The judgments that typically survive a Delphi procedure may not be the best judgments but represent, rather, the position of minimum compromise.  As a result, the surviving judgments may lack the significance that extreme or conflicting positions may possess.  Further, the "raw data," "facts," or "simple observables" of the empiricist on deeper analysis have always proved to be exceedingly complex and hence further divisible into other entities thought to be indivisible or simple, ad infinitum.

More troublesome still is the almost extreme and unreflective reliance on agreement as the sole or major principle for producing information, and even truth, out of raw data.  Agreement may stifle conflict and debate when they are needed most, and its cost can be prohibitive.  As a result, Lockean systems are best suited for working on well-structured probable situations for which there exists a strong consensual position on the nature of the problem situation.  If these conditions or assumptions cannot be met or justified by the decision-maker -- for example, if it seems too risky to base projections of what the future will be like on the judgments of experts, no matter how strong the agreement between them -- then some alternate system or inquiry  may be called for, as in the previous case of the Leibnizian inquirer.

Kantian inquiry
The last two sections have illustrated the difficulties that arise from emphasizing one of the components of a tightly coupled system of inquiry to the detriment of the other components.  Leibnizian inquiry emphasizes theory to the detriment of data and Lockean inquiry emphasizes data to the detriment of theory.  When translate into practice, what often results is highly sophisticated models with little or no concern for the difficult problems associated with the collection of data or the seemingly endless proliferation of data with little regard for the dictates of currently existing models.

The recent controversy surrounding the attempts of Jay Forrester and Dennis Meadows, at M.I.T., to build a "world model" is a good illustration of the strong differences between these two points of view.  The work of Forrester and Meadows represents a almost pure Leibnizian approach to the modeling of large complicated systems.  Their model is, in effect, data independent.  One can criticize the model on pure Leibnizian grounds, e.g., whether the internal theory and structure of the model are sound with respect to current economic and social theory, and some of the critics have chosen to do this.  However, it would seem that more often than not the critics have chosen to offer a Lockean critique, i.e., that some other way, say, using accurate statistical data, is a better way to build a sound forecast model of the world.  Although this is a legitimate method of criticism, to a large extent it only further exacerbates the differences between the two approaches.  Hence it misses the real point, which is not whether the Forrester-Meadows approach is the correct Leibnizian approach, or whether there is a correct Lockean approach, but rather whether any Leibnizian or Lockean approach acting independently of the other could ever possibly be "correct."

Forrester and Meadows seek to validate their approach through the robustness and richness of their model, and their Lockean critics attempt to establish the validity of their approach through the priority and "regularity" of the statistical data to which they appeal.  If the debate proves anything, it raises the serious question as to whether an advanced society can continue to rely on purely Leibnizian or Lockean efforts for its planning. To really evaluate the relative merits of separate Leibnizian or Lockean inquirers, it is necessary to go to a philosophy that incorporates both, such as the Kantian inquirer.

The sense of Kantian inquiry can be rather quickly grasped form the general characteristic that ruth is synthetic; i.e., the truth content of a system is not located in either its theoretical or its empirical components, but in both.

A corollary is that neither the data input nor the theory has priority.  Theories or general propositions are built up form data, and in this sense theories are dependent on data, but data cannot be collected without the prior assumption of some theory of data collection (a theory of "how to make observations," "what to observe," etc.), and in this sense data depend on theories.  Theory and data are inseparable.

An important feature of Kantian inquiry is that for any problem, one must build at least two alternative representations or models.  The hope is that out of these alternate representations, or fact nets, of a decision-maker's or client's problem, there will be one that is "best" for representing the problem.  The defect of Leibnizian and Lockean inquiry is that they give only one view of the problem.  Kantian inquiry attempts to give many explicit views.  The guarantor of such systems is the degree of fit or match between the underlying theory (theoretical predictions) and the data collected under the presumption of that theory.

Kantian inquiry places such heavy emphasis on alternate models because, in dealing with problems such as the nature of the future, the real problem is how to get as many perspectives as possible on the nature of the subject problem.  Problems like the future cannot be formulated and solved via a single well-structured approach.  In dealing with the future, we are not dealing with the concrete realities of human existence, but, if only in part, with hopes, dreams, plans, and aspirations.  Since different men rarely share the same aspirations, it seems that the best way to "analyze" aspirations is to compare as many of them as we can.  If the future is 99 percent aspiration or plan, it would seem that the best approach is to draw forth explicitly as many different aspirations or plans for the future as possible.  In short, we want to examine as many different alternate futures as we can.

In the field of technological forecasting, normative forecasting, planning programming budgeting systems (PPBS), and cost-effectiveness or cost-benefit analysis are all examples of Kantian inquiry, although at such a low level as to be almost more Leibnizian than Kantian in nature.  The Kantian element these approaches share is the fact that they are all concerned with alternate paths or methods of getting from a present state to a future state characterized by certain objectives, needs, or goals (or vice versa).  When these various planning vehicles have failed, it has often been a problem of unclear or fuzzy objectives or poor compatibility among data, models, and objectives.  Furthermore, the systems are usually applied with a questionable and implicit Leibnizian assumption that all benefit or effectiveness measures can be expressed in dollars.

In recent years, there have been a number of Delphi studies that more actively take on the characteristics of Kantian inquiry.  These differ fundamentally from the original Delphis, which were strongly Lockean in orientation.  The initial Delphis were characterized by a strong emphasis on the use of consensus by a group of "experts" as the means to converge on a single model or position on some issue.  In contrast, the explicit purpose of a Kantian Delphi is to elicit alternatives on which to base a comprehensive overview of the issue.  In terms of communication processes, although a "consensus" or Lockean Delphi is better suited to setting up a communications structure among an already informed group that possesses the same general core of knowledge, a Kantian or "contributory" Delphi attempts to design a structure that allows many "informed" individuals in different disciplines or specialties to contribute information or judgments to a problem area that is much broader in scope than the knowledge that any on of the individuals possesses.

This type of Delphi has been applied to conceptualizing such problems as: (1) defining a structural model for material flows in the steel industry; (2) examining the present and the potential role of the mentally retarded in society; (3) forecasting the future characteristics of recreation and leisure; and (4) examining the past history of the internal combustion engine for a clue to significant events possibly affecting its future.  Although all of these Delphis had specific forecasting objectives, the problems are so broad that the objectives could not be achieved if the parties to the Delphi were from the same specialized interest group.  (For example, educators, psychiatrists, parents, and teachers all have different and valid perspectives to contribute to the definition of the "problem" of the mentally retarded.)  Thus, the goal, at least in the initial stages, is not to reach consensus on a single definition but rather to elicit many divers points of view and potential aspects of the problem.  In essence, the objective is to establish how to fit the pieces of a jigsaw together, and even to determine if it is one or many puzzles.

Kantian inquiry is best suited to problems that are inherently ill-structured; i.e., the kinds of problems that are inherently difficult to formulate in pure Leibnizian or Lockean terms because their nature does not admit of a clear consensus or a simple analytic attack.  On the other hand, the Kantian inquiry is not applicable to the kinds of problems that admit of a single clear formulation because here the proliferation of alternate models may be too costly or time consuming.  Kantian inquiry may also overwhelm those who are used to "the single best model" approach to any problem.  Of course, this in itself is not necessarily bad if it helps to teach those who hold this belief that there are some kinds of problems for which there is no one best approach.  Social problems inherently seem to be of this kind and thus to call for a Kantian approach.  The concept of "technology assessment" as a vehicle for determining the relationships between technology and social consequences would also seem to imply the necessity of at least a Kantian approach.  Many efforts labeled as assessments have proved inadequate because they were conducted as Leibnizian or Lockean inquiries.

Hegelian inquiry
The fourth in our spectrum of inquiry systems is the Hegelian, or dialectical, inquiry.  Its basic idea is that truth is conflictual, that is, the truth content of a system is the result of a highly complicated process that depends on the existence of a plan and a diametrically opposed counterplan.  The plan and the counterplan represent strongly divergent and opposing conceptions of the whole system.. Their function is to engage each other in an unremitting debate over the "true" nature of the whole system, in order to draw forth a new plan that will hopefully reconcile (synthesize, encompass) the plan and the counterplan.  A corollary to this is that by itself the data input sector is totally meaningless and only becomes meaningful -- i.e., "information" -- by being coupled to the plan and the counterplan.

Thus, Hegelian inquiry systems are the epitome of conflictual, synthetic systems.  They build at least two, completely antithetical, representations of any problem.  Hegelian inquiry starts by identifying or creating two strongly opposing Leibnizian models of a problem that constitute the contrary underlying assumptions regarding the problem's theoretical nature.  Both of these Leibnizian representations are then applied to the same Lockean data set in order to demonstrate that the same data set can be used to support either theoretical model.  The point is that data are not information; information results from the interpretation of data.  It is intended that out of a dialectical confrontation between opposing interpretations (e.g., the opposing "expert" views of a situation), the underlying assumptions of both Leibnizian models (or opposing policy experts) will be brought to the surface for conscious examination by the decision-maker, who is dependent upon his experts for advice.  It is also hoped that as a result of witnessing the dialectical confrontation between experts or models, the decision-maker will be in a better position to form his own view (build his own model or become his own expert) on the problem that is a "creative synthesis" of the two opposing views.  Whereas in the Lockean inquiry the guarantor is agreement, in the Hegelian it is intense conflict -- the presumption that conflict will expose the assumptions underlying an expert's point of view that are often obscured precisely because of the agreement between experts.

Hegelian inquiry is best suited for studying ill-structured problems.  These are the problems that, precisely because of their poor structure, will produce intense debate over their "true" nature.  Conversely, it is not recommended for well-structured, clear-cut problems because here conflict may be a time-consuming nuisance.

Except for the policy Delphi concept of Turoff, the use of conflict as a methodology is conspicuously absent in the field of technological forecasting.  In the "policy Delphi" the communication process is designed to produce the best pro or con arguments underlying various policy alternatives or resource-allocation alternatives.  In a non-Delphi (face-to-face) mode one of the most interesting applications can be found in the activity of corporate or strategic planning.  In an important case study, Richard Mason literally pioneered the development of what may be termed the dialectical inquiring systems (DIS).

The situation encounter by Mason was one in which the nature of the problem prevented traditional well-structured technical approaches to planning (Leibnizian and Lockean) from being used.  Mason studied a company where two strongly opposing groups of top executives had almost completely contrary views about the fundamental nature and management of their organization.  faced with a crucial decision concerning the company's future, each group offered fundamentally differing plans as to how to cope with the situation.  Neither plan could be proved or "checked out" by performing any technical study, since each plan rested on a host of assumptions, many of them unstated, that could probably never be verified in their entirety even if sufficient time had been available.  Indeed, if the executives wanted to be around in the future to check on how well their assumptions turned out, they had to make a decision in the present.  It was at this point that the company agreed to let Mason try the DIS.

After careful study and extensive interview with both sides, Mason assembled both groups of executives and made the following presentation: first, he laid out side by side on opposite halves of a display board what he took to be the underlying assumptions on which the two groups were divided.  Thus, for every assumption of the one side there was an opposing assumption for the other side.  Next, Mason took a typical set of characteristic operating data on the present state of the company (profit, rate of return on investment, etc.) and showed that every piece of data could be used to support either the plan or the counterplan; i.e., there was an interpretation of the data that was consistent with both plans.  Hence, the real debate was never really over the data, as the executives had previously thought, but over the underlying assumptions.  Finally, as a result of witnessing this, both groups were asked if they, not Mason, could now formulate a new plan that encompassed their old plans.  fortunately they could, and because of the intense and heated debate that took place, both groups felt they had achieved a better examination of their proposed course of action.

Of Course, such a procedure does not guarantee an optimal solution.  But then, the DIS is most applicable to those situations in which the problem cannot be formulated in pure Leibnizian terms for which a unique, optimal solution can be derived.  DIS is most appropriate for precisely those situations in which there is no better tool to relay on than the opinions of opposing experts.  Where the future is 99 percent opinion and assumptions DIS may be most apt.

The DIS and policy Delphis differ fundamentally from other techniques and procedures that make use of conflict.  In an ordinary courtroom debate, for instance, both sides are free to introduce whatever supporting data and opposing arguments they wish.  Thus, the two are confounded.  In a DIS or a policy Delphi the opposing arguments are kept strictly apart from the data so that the crucial function of the opposing arguments can be explicitly demonstrated.  This introduces an element of artificiality that real debates do not have, but then it also introduces a strong element of structure and clarity that makes this use of conflict much more controlled and systematic.  In essence, the Hegelian inquiry process dictates a conceptual communication structure that relates the conflict to the data and the objectives.  Under this conception of inquiry, conflict is no longer antithetical to Western science's preoccupation with objectivity; indeed, conflict actually serves objectivity in this case.  This perhaps will be puzzling to those who have been brought up on the idea that objectivity is that upon which men agree and not on what they disagree.  Although the Hegelian inquirer does not always lead to a new agreement, or a new plan, when it does the agreement is likely to be stronger.

The Singerian system of inquiry
The most complicated of the inquirers discussed here, and hence the most difficult to describe fully, is based on the philosophy of the early 20th century American pragmatist, Edgar Singer.  Its main features are as follows: 'Truth is pragmatic; that is, the truth content of a system is relative to the overall goals and objectives of the inquiry.  A model of a system is teleological or explicitly goal-oriented, in the sense that the truth of the model is measured with respect to its ability to articulate certain systems objectives, to create several alternate means for securing these objectives, and finally, at the "end" of the inquiry, to specify new goals that remain to be accomplished by some future inquiry.  Singerian inquirers thus never give final answers to any question, although at any point they seek to give a refined, specific response.

As a corollary, Singerian inquiry systems are the most strongly coupled of all the inquirers.  No single aspect of the system has any fundamental priority over any of the other aspects.  The system forms an inseparable whole.  Singerian inquiry takes holistic thinking so seriously that it constantly attempts to sweep in new variables and additional components to broaden the base of concern.  For example, it is an explicit postulate of Singerian inquiry that the system designer is a fundamental part of the system, and as a result his psychology and sociology must be explicitly considered as one of the system components.

Singerian inquirers are the epitome of synthetic, multimodel, interdisciplinary systems.  In effect, Singerian inquiry constitutes a theory about all the other inquirers (Leibnizian, Lockean, Kantian, Hegelian), and forms a theory abut how to manage their application.

Singerian inquiry systems contain some rather distinctive features that none of the others possess.  One is that they speak almost exclusively in the language of commands; for example, "Take this model of the system as the true mode." The point is that all of the models, laws, and facts of science are only approximations.  The "hard facts: and "firm laws" of science are only "facts" and "laws" if we are willing to accept certain strong assumptions about the nature of the reality underlying the measurement of the facts and the operation of the laws.  The thing that serves to legitimize these assumptions in the command, in whatever form it is expressed, to take them seriously {"Take this as the true model underlying the phenomenon in question so that with this model as a background we can do such and such experiments").  Thus, for example, the Bohr model of the atom is not a "factually real description of the atom" but if we regard it as such we can perform certain experiments and make certain theoretical predictions that we would be unable to do without the model.  What Singerian inquirers do is to draw these hidden commands out of every system so that the analyst is hopefully in a better position to choose his commands carefully.

Singerian inquiry also greatly expands on the potential set of system designers and users.  In the extreme, the set is broadened to include all of mankind, since in an age of larger and larger systems nearly everyone is affected by or affects every other system.  Singerian inquirers attempt to base their forecast of the future on the projections of as many diverse discipline, professions, and personalities as possible.

As far as we know, Singerian inquiry is virtually absent form the field of technological forecasting and assessment.  However, the implication of Singerian inquiry for technological forecasting is that the supposed "fundamental polarity of exploratory and normative technological forecasting" completely breaks down.  According to conventional wisdom, "exploratory technological forecasting starts from today's assured basis of knowledge and is oriented toward the future, whereas normative technology forecasting first assesses future goals, needs, desires, missions, etc., and works backward to the present.:  (Janfsch; see "Recommended Reading."  However comforting this sounds, it ignores the basic Singerian point that every description of the present (“today's assured basis of knowledge") is based on some normative conception of the future (i.e., "future goals, needs, desires, missions, etc."). In Singerian terms, it is incredibly naive to take as "fundamental polarities" that which fundamentally interacts.  Our normative plans for the future are idealized plans for expanding our knowledge of "what is known in the present."  One of the reasons why man has always been interested in the future is that he has always been satisfied with that which he has and knows in the present.  Our plans for the future express what we wish the present were like.

The strength of Singerian inquiry is that it gives the broadest possible modeling of any inquirer on any problem.  The weakness is the potentially prohibitive costs involved in such comprehensive modeling efforts.  However, given the increased fear and concern with our environment, we may no longer have the choice but to pay the price.  We may no longer be able to afford the continued "luxury" of building large scale Leibnizian and Lockean technological models devoid of the serious and explicit ethical considerations that can be handled with Singerian inquiry.

A look at structure
Having now examined five philosophies of forecasting, we will turn briefly to a consideration of what might be called its structure (and assessment).
1. Feasible technological developments.  Feasible usually means, in this context, technically feasible if the "required " resources are invested or available.
2. Potential applications.  This is any possible application of the previous technological developments without regard to their "good" or "bad" values.
3. Significant applications.  This is some subset of "all" potential applications or a transformation to some set that is significant to the study's intent.
4. Potential consequences.  Any consequences, "good" or "bad," that may affect opinions of scenarios about the future, or our interpretation of the past.
5. Policy or resource allocation issues.  The decision questions under examination or arising as a result of observing potential consequences.
6. Potential resolutions of issues.  The controls that can be imposed to affect the likelihood of various developments, applications, and consequences.

In practice, most technological forecasting and assessment studies focus attention on one of these six categories and treat   others with various degrees of implicitness or explicitness.  We are in a situation today very reminiscent of the blind man and the elephant.  It is common to find engineering forecasting studies that focus only on the first or second elements -- developments and applications -- with little reference to the other items.  In contrast, those who look at the assessment elephant from the view of the social sciences usually focus on the consequences and policy issues.

Frequently the new technological assessment efforts are looked upon as an entirely different breed of animal from the "classical: technological forecasting used for organizational planning purposes.  It is interesting, therefore, that one observation we can make explicitly form this structure is that the only evident distinction between the two is in how we define the scope of the "potential consequences.:  In forecasting, we are concerned with the effects on the organization (profits, markets, mission objectives, etc.); and in the assessment the effects of concern are those on society (changes in lifestyle, job markets, education, pollution, etc.).

There are two considerations that considerably complicate the deceivingly simple structure for technological forecasting and assessment.  The first is the problem of "enumeration" -- how does one attempt to ensure that all relevant pieces of information are included in the analysis?  The morphological approach to this is the process of finding a model for classifying "all" items within a category into some finite set of subcategories that span the region of interest.  In many case these subcategories are tied to specific ranges of physical parameters such as velocity, frequency of radiation, weight, etc.  Although this approach works well when talking about developments or applications, immediate difficulties or disagreements arise when one moves into the area of consequences or policy.

The second aspect of complication lies in attempting to describe the interactions, interrelationships, and causal effect among those various enumerated items.  Our view of the future is dependent upon our view of the present and the resulting view of the past.  Given ten events about the future there are about ten million relationships that could, in principle, be described among this small event set.  Many of the techniques if forecasting are merely attempts to define a less involved and approximate structure that is sufficient for picking out the significant interactions in any set of items.  These approaches fall broadly into two general categories: matrix and network representations.  Some of the names under which these two approaches are often disguised are cross impact, cross support, management matrices, relevance trees, decision networks or trees, and patterns.

When a well understood structure exists that is fairly sparse with respect to interactions among the items, then a network or tree structure is often used.  When the structure is not well understood or not sparse, various matrix methods are usually employed for defining the structure.  If a good morphological set has been defined, the techniques for defining these relationships may be applied to the elements of the morphological representation, as opposed to the original information items.  Since there are an unlimited number of ways we can model the future, there exists a rich an growing literature on these morphological and impact or relationship techniques.  for the limited objective of this discussion, a concept of the inquiry process associated with each step in the technological forecasting and assessment cycle should be sufficient to provide the reader with a perspective for evaluating these various techniques.

The process of delineating and examining technological developments and applications can be handled by setting up a Leibnizian or Lockean inquirer that utilizes various implicit future scenarios and representations of past as the raw data input.  The problem of determining "significant" applications and the resulting potential consequences dictates at least the use of a Lockean inquirer and possibly a Kantian inquirer.  Especially when the problem is more of an assessment than a forecast, the Kantian approach should be mandatory for this part of the cycle.  In the area of policy and resource allocation, either a Kantian or Hegelian process would seem to be appropriate.  At this point most study efforts usually terminate.  However, the forecasting process is best viewed as a continuous cycle with two important feedback loops: the over inquiry process should cause us both to examine the past for its possible reinterpretation, and to reconceive our conceptualization of the future.  The two of these taken together represent a Singerian process that ties all the other elements of the system together in a continuous reflective cyclic process.

When technology forecasting and assessment are view form this perspective, the process of studying the future becomes inseparable form the process of studying the past.  A good forecaster should therefore be a good historian.

Finally ...
In conclusion, we would point out that what separates science from mythology is not the subject matter of an inquiry but the approach.  Something is a science if it can show (1) what that something needs to control, and (2) how to control it so that someone can study it in a controlled and systematic or scientific way.  In the field of technological forecasting we are just beginning to be aware of the first part, i.e., that the number of things we need to control (study) in order to make forecasts is indeed large.  At the very minimum we need not only sweep in the things that the physical and social science study, but those that the humanities study as well, such as ethics.

In the end, it is the philosophical ability to be self-reflective that separates science from mythology.  Self-reflection implies a realization that as much as our inquiry models describe and represent reality, they also describe and represent us, our psychology.  thus, for example, reflection points out that the mathematical type (the Leibnizian analyst) has an incessant need to reduce every problem to a mathematical one, even where it is not appropriate of efficient; the realist (Lockean) exclusively associates reality with facts or hard data even where the  data are limited and confining; the idealist (Kantian) associates reality with possibilities even where they are not feasible; the pragmatist (Singerian) associates reality with the feasible or the do-able, even when it is not worth doing; and the conflictual (Hegelian) restricts reality to that which survives a strong debate even where a debate is not called for.  The difference between science and mythology is that the former, unlike the later, attempts to study itself -- to raise to consciousness its underlying premises and psychology.  In short, a scientist understands the philosophy underlying what he is doing.  Applying "scientific" methods without this understanding is the application of a methodology.

A more detailed version of this paper will be found in the Journal of Technological Forecasting and Social Change, vol. 5. no 1. fall 1973.
 
 

Philosophical role playing in the executive suite

In any discussion involving such topics as planning and assessment it is not difficult to relate the statements and questions raised to the various philosophies of inquiry.  Consider, for example, a group of manager discussing a prospective project.  A Lockean manager might well begin with the typical intuitive assertions: "Give me these particular people and I'll be able to do the job." If Leibniz were in the room he would probably respond with: "You do the job with the people you have!" Underlying this reply is a model that a certain number of people working a certain number of hours will be able to do a certain job, and this is independent of the data (in this case of who do the work). In contrast, Kant, who is interested in objectives, would probably ask: "Why do you want to do the job?" and Hegel would pose a significant variation of Kant's question: "What are the advantages of not doing the job?" Hegel wants to be sure that the opposing view is recognized and that we might not be better off by not doing the job. Finally, there is Singer, who, unless he happens to be the boss, is usually the person most prone to getting fired.  Because Singer tends to reflect on what is taking place and seek out the hidden assumptions or underlying psychology, he has a tendency to discover what most individuals have subconsciously agreed not to discuss.  In this case he might very well wish to broaden the discussion by asking: "Why do you have the people you have if they cannot do the job:"

Ian I Mitroff is an associate professor in the Graduate School of Business and the Interdisciplinary Doctoral Program in Information Science at the University of Pittsburgh.  He also holds two research appointments, in the Philosophy of Science Center and in the Learning, Research, and Development Center.  His Ph.D. is in engineering science with a minor in the philosophy of science from the University of California, Berkeley.  His current research includes the design of philosophically based information systems.  He has also recently completed a major sociology of science study of the Apollo moon scientists.
 

Murray Turoff is currently with the Systems Evaluation Division in the Office of Emergency Preparedness of the Executive Office of the President.  He has been associated with the Institute for Defense Analysis and IBM.  His principal area of professional interest are Delphi design, information systems, modeling, simulation, gaming, and technological forecasting.  Dr. Turoff currently teaches as course in technological forecasting at the American University.  He received the PH.D. in physics from Brandeis University and the B.A. degree in mathematics and physics from the University of California at Berkeley.  He has, however, been working primarily in operations research and computer applications since 1964, and is the author of a number of papers.

IEEE Spectrum March 1973 p62-71